Unnamed repository; edit this file 'description' to name the repository.
1
2
3
4
5
6
7
8
9
((regex_lit) @injection.content
 (#set! injection.language "regex"))
id='n54' href='#n54'>54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354
use crate::graphemes::{nth_next_grapheme_boundary, nth_prev_grapheme_boundary, RopeGraphemes};
use crate::syntax::LOADER;
use crate::{ChangeSet, Diagnostic, Position, Range, Rope, RopeSlice, Selection, Syntax};
use anyhow::Error;

/// A state represents the current editor state of a single buffer.
#[derive(Clone)]
pub struct State {
    // TODO: fields should be private but we need to refactor commands.rs first
    pub doc: Rope,
    pub selection: Selection,
}

#[derive(Copy, Clone, PartialEq, Eq)]
pub enum Direction {
    Forward,
    Backward,
}
#[derive(Copy, Clone, PartialEq, Eq)]
pub enum Granularity {
    Character,
    Word,
    Line,
    // LineBoundary
}

impl State {
    #[must_use]
    pub fn new(doc: Rope) -> Self {
        Self {
            doc,
            selection: Selection::single(0, 0),
        }
    }

    // TODO: doc/selection accessors

    // TODO: be able to take either Rope or RopeSlice
    #[inline]
    pub fn doc(&self) -> &Rope {
        &self.doc
    }

    #[inline]
    pub fn selection(&self) -> &Selection {
        &self.selection
    }

    // pub fn doc<R>(&self, range: R) -> RopeSlice
    // where
    //     R: std::ops::RangeBounds<usize>,
    // {
    //     self.doc.slice(range)
    // }

    // update/transact:
    // update(desc) => transaction ?  transaction.doc() for applied doc
    // transaction.apply(doc)
    // doc.transact(fn -> ... end)

    // replaceSelection (transaction that replaces selection)
    // changeByRange
    // changes
    // slice
    //
    // getters:
    // tabSize
    // indentUnit
    // languageDataAt()
    //
    // config:
    // indentation
    // tabSize
    // lineUnit
    // syntax
    // foldable
    // changeFilter/transactionFilter

    // TODO: move that accepts a boundary matcher fn/list, we keep incrementing until we hit
    // a boundary

    // TODO: edits, does each keypress trigger a full command? I guess it's adding to the same
    // transaction
    // There should be three pieces of the state: current transaction, the original doc, "preview"
    // of the new state.
    // 1. apply the newly generated keypress as a transaction
    // 2. compose onto a ongoing transaction
    // 3. on insert mode leave, that transaction gets stored into undo history

    pub fn move_pos(
        &self,
        pos: usize,
        dir: Direction,
        granularity: Granularity,
        count: usize,
    ) -> usize {
        let text = &self.doc;
        match (dir, granularity) {
            (Direction::Backward, Granularity::Character) => {
                // Clamp to line
                let line = text.char_to_line(pos);
                let start = text.line_to_char(line);
                std::cmp::max(
                    nth_prev_grapheme_boundary(&text.slice(..), pos, count),
                    start,
                )
            }
            (Direction::Forward, Granularity::Character) => {
                // Clamp to line
                let line = text.char_to_line(pos);
                // Line end is pos at the start of next line - 1
                // subtract another 1 because the line ends with \n
                let end = text.line_to_char(line + 1).saturating_sub(2);
                std::cmp::min(nth_next_grapheme_boundary(&text.slice(..), pos, count), end)
            }
            (Direction::Forward, Granularity::Word) => {
                Self::move_next_word_start(&text.slice(..), pos)
            }
            (Direction::Backward, Granularity::Word) => {
                Self::move_prev_word_start(&text.slice(..), pos)
            }
            (_, Granularity::Line) => move_vertically(&text.slice(..), dir, pos, count),
            _ => pos,
        }
    }

    pub fn move_next_word_start(slice: &RopeSlice, mut pos: usize) -> usize {
        // TODO: confirm it's fine without using graphemes, I think it should be
        let ch = slice.char(pos);
        let next = slice.char(pos.saturating_add(1));
        if categorize(ch) != categorize(next) {
            pos += 1;
        }

        // refetch
        let ch = slice.char(pos);

        if is_word(ch) {
            skip_over_next(slice, &mut pos, is_word);
        } else if ch.is_ascii_punctuation() {
            skip_over_next(slice, &mut pos, |ch| ch.is_ascii_punctuation());
        }

        // TODO: don't include newline?
        skip_over_next(slice, &mut pos, |ch| ch.is_ascii_whitespace());

        pos
    }

    pub fn move_prev_word_start(slice: &RopeSlice, mut pos: usize) -> usize {
        // TODO: confirm it's fine without using graphemes, I think it should be
        let ch = slice.char(pos);
        let prev = slice.char(pos.saturating_sub(1)); // TODO: just return original pos if at start

        if categorize(ch) != categorize(prev) {
            pos -= 1;
        }

        // TODO: skip while eol

        // TODO: don't include newline?
        skip_over_prev(slice, &mut pos, |ch| ch.is_ascii_whitespace());

        // refetch
        let ch = slice.char(pos);

        if is_word(ch) {
            skip_over_prev(slice, &mut pos, is_word);
        } else if ch.is_ascii_punctuation() {
            skip_over_prev(slice, &mut pos, |ch| ch.is_ascii_punctuation());
        }

        pos.saturating_add(1)
    }

    pub fn move_next_word_end(slice: &RopeSlice, mut pos: usize, _count: usize) -> usize {
        // TODO: confirm it's fine without using graphemes, I think it should be
        let ch = slice.char(pos);
        let next = slice.char(pos.saturating_add(1));
        if categorize(ch) != categorize(next) {
            pos += 1;
        }

        // TODO: don't include newline?
        skip_over_next(slice, &mut pos, |ch| ch.is_ascii_whitespace());

        // refetch
        let ch = slice.char(pos);

        if is_word(ch) {
            skip_over_next(slice, &mut pos, is_word);
        } else if ch.is_ascii_punctuation() {
            skip_over_next(slice, &mut pos, |ch| ch.is_ascii_punctuation());
        }

        pos.saturating_sub(1)
    }

    pub fn move_selection(
        &self,
        dir: Direction,
        granularity: Granularity,
        count: usize,
    ) -> Selection {
        // move all selections according to normal cursor move semantics by collapsing it
        // into cursors and moving them vertically

        self.selection.transform(|range| {
            let pos = self.move_pos(range.head, dir, granularity, count);
            Range::new(pos, pos)
        })
    }

    pub fn extend_selection(
        &self,
        dir: Direction,
        granularity: Granularity,
        count: usize,
    ) -> Selection {
        self.selection.transform(|range| {
            let pos = self.move_pos(range.head, dir, granularity, count);
            Range::new(range.anchor, pos)
        })
    }
}

/// Convert a character index to (line, column) coordinates.
pub fn coords_at_pos(text: &RopeSlice, pos: usize) -> Position {
    let line = text.char_to_line(pos);
    let line_start = text.line_to_char(line);
    let col = text.slice(line_start..pos).len_chars();
    Position::new(line, col)
}

/// Convert (line, column) coordinates to a character index.
pub fn pos_at_coords(text: &RopeSlice, coords: Position) -> usize {
    let Position { row, col } = coords;
    let line_start = text.line_to_char(row);
    nth_next_grapheme_boundary(text, line_start, col)
}

fn move_vertically(text: &RopeSlice, dir: Direction, pos: usize, count: usize) -> usize {
    let Position { row, col } = coords_at_pos(text, pos);

    let new_line = match dir {
        Direction::Backward => row.saturating_sub(count),
        Direction::Forward => std::cmp::min(row.saturating_add(count), text.len_lines() - 1),
    };

    // convert to 0-indexed, subtract another 1 because len_chars() counts \n
    let new_line_len = text.line(new_line).len_chars().saturating_sub(2);

    let new_col = if new_line_len < col {
        // TODO: preserve horiz here
        new_line_len
    } else {
        col
    };

    pos_at_coords(text, Position::new(new_line, new_col))
}

// used for by-word movement

fn is_word(ch: char) -> bool {
    ch.is_alphanumeric() || ch == '_'
}

#[derive(Debug, Eq, PartialEq)]
enum Category {
    Whitespace,
    EOL,
    Word,
    Punctuation,
}
fn categorize(ch: char) -> Category {
    if ch == '\n' {
        Category::EOL
    } else if ch.is_ascii_whitespace() {
        Category::Whitespace
    } else if ch.is_ascii_punctuation() {
        Category::Punctuation
    } else if ch.is_ascii_alphanumeric() {
        Category::Word
    } else {
        unreachable!()
    }
}

fn skip_over_next<F>(slice: &RopeSlice, pos: &mut usize, fun: F)
where
    F: Fn(char) -> bool,
{
    let mut chars = slice.chars_at(*pos);

    for ch in chars {
        if !fun(ch) {
            break;
        }
        *pos += 1;
    }
}

fn skip_over_prev<F>(slice: &RopeSlice, pos: &mut usize, fun: F)
where
    F: Fn(char) -> bool,
{
    // need to +1 so that prev() includes current char
    let mut chars = slice.chars_at(*pos + 1);

    while let Some(ch) = chars.prev() {
        if !fun(ch) {
            break;
        }
        *pos -= 1;
    }
}

#[cfg(test)]
mod test {
    use super::*;

    #[test]
    fn test_coords_at_pos() {
        let text = Rope::from("ḧëḷḷö\nẅöṛḷḋ");
        assert_eq!(coords_at_pos(&text.slice(..), 0), (0, 0).into());
        assert_eq!(coords_at_pos(&text.slice(..), 5), (0, 5).into()); // position on \n
        assert_eq!(coords_at_pos(&text.slice(..), 6), (1, 0).into()); // position on w
        assert_eq!(coords_at_pos(&text.slice(..), 7), (1, 1).into()); // position on o
        assert_eq!(coords_at_pos(&text.slice(..), 10), (1, 4).into()); // position on d
    }

    #[test]
    fn test_pos_at_coords() {
        let text = Rope::from("ḧëḷḷö\nẅöṛḷḋ");
        assert_eq!(pos_at_coords(&text.slice(..), (0, 0).into()), 0);
        assert_eq!(pos_at_coords(&text.slice(..), (0, 5).into()), 5); // position on \n
        assert_eq!(pos_at_coords(&text.slice(..), (1, 0).into()), 6); // position on w
        assert_eq!(pos_at_coords(&text.slice(..), (1, 1).into()), 7); // position on o
        assert_eq!(pos_at_coords(&text.slice(..), (1, 4).into()), 10); // position on d
    }

    #[test]
    fn test_vertical_move() {
        let text = Rope::from("abcd\nefg\nwrs");
        let pos = pos_at_coords(&text.slice(..), (0, 4).into());
        let slice = text.slice(..);

        assert_eq!(
            coords_at_pos(&slice, move_vertically(&slice, Direction::Forward, pos, 1)),
            (1, 2).into()
        );
    }
}